On Linux, is there a way to embed version information into an ELF binary? I would like to embed this info at compile time so it can then be extract it using a script later. A hackish way would be to plant something that can be extracted using the strings command. Is there a more conventional method, similar to how Visual Studio plant version info for Windows DLLs (note version tab in DLL properties)?
One way to do it if using cvs or subversion is to have a special id string formatted specially in your source file. Then add a pre-commit hook to cvs or svn that updates that special variable with the new version of the file when a change is committed. Then, when the binary is built, you can use ident to extract that indformation. For example:
Add something like this to your cpp file:
static char fileid[] = "$Id: fname.cc,v 1.124 2010/07/21 06:38:45 author Exp $";
And running ident (which you can find by installing rcs) on the program should show the info about the files that have an id string in them.
ident program
program:
$Id: fname.cc,v 1.124 2010/07/21 06:38:45 author Exp $
Note As people have mentioned in the comments this technique is archaic. Having the source control system automatically change your source code is ugly and the fact that source control has improved since the days when cvs was the only option means that you can find a better way to achieve the same goals.
To extend the #sashang answer, while avoiding the "$Id:$" issues mentioned by #cdunn2001, ...
You can add a file "version_info.h" to your project that has only:
#define VERSION_MAJOR "1"
#define VERSION_MINOR "0"
#define VERSION_PATCH "0"
#define VERSION_BUILD "0"
And in your main.c file have the line:
static char version[] = VERSION_MAJOR "." VERSION_MINOR "." VERSION_PATCH "." VERSION_BUILD;
static char timestamp[] = __DATE__ " " __TIME__;
(or however you want to use these values in your program)
Then set up a pre-build step which reads the version_info.h file, bumps the numbers appropriately, and writes it back out again. A daily build would just bump the VERSION_BUILD number, while a more serious release would bump other numbers.
If your makefile lists this on your object's prerequisite list, then the build will recompile what it needs to.
The Intel Fortran and C++ compilers can certainly do this, use the -sox option. So, yes there is a way. I don't know of any widespread convention for embedding such information in a binary and I generally use Emacs in hexl-mode for reading the embedded information, which is quite hackish.
'-sox' also embeds the compiler options used to build an executable, which is very useful.
If you declare a variable called program_version or similar you can find out at which address the variable is stored at and then proceed to extract its value. E.g.
objdump -t --demangle /tmp/libtest.so | grep program_version
0000000000600a24 g O .data 0000000000000004 program_version
tells me that program_version resides at address 0000000000600a24 and is of size 4. Then just read the value at that address in the file.
Or you could just write a simple program that links the library in questions and prints the version, defined either as an exported variable or a function.
Related
I am following some Linux instructions on OS X and am stuck on one line:
strip -N main my_file.o
The OS X version of strip doesn't have an -N option and I've read through the man page twice but am just not sure how to do this.
So how do I strip a single name from the symbol table on OS X?
As you say, the OSX version of strip doesn't allow this; the only way therefore is to limit its visibility in code using this on the declaration:
__attribute__((visibility("hidden"))) void MyFunction1();
Alternatively you could compile with -fvisibility=hidden and use "default" in the above __attribute__ to only expose the symbols you want.
This is a better approach anyway, as it does not require an external build step.
Note: I have found this doesn't work as expected when Objective-C code is introduced into the code base...
Reference
I compile twice the same .c and .h files and get object files with the same size but different md5sums.
Here is the only difference from objdump -d:
1) cpcidskephemerissegment.o: file format elf64-x86-64
Disassembly of section .text:
0000000000000000 <_ZN68_GLOBAL__N_sdk_segment_cpcidskephemerissegment.cpp_00000000_B8B9E66611MinFunctionEii>:
2) cpcidskephemerissegment.o: file format elf64-x86-64
Disassembly of section .text:
0000000000000000 <_ZN68_GLOBAL__N_sdk_segment_cpcidskephemerissegment.cpp_00000000_8B65537811MinFunctionEii>:
What can be the reason? Thanks!
I guess, the compiler didn't know how to name this namespace and used path to the source file plus some random number.
The compiler must guarantee that a symbol in unnamed namespace does not conflict with any other symbol in your program. By default this is achieved by taking full filename of the source, and appending a random hash value to it (it's legal to compile the same source twice (e.g. with different macros) and link the two objects into a single program, and the unnamed namespace symbols must still be distinct, so using just the source filename without the seed is not enough).
If you know that you are not linking the same source file more than once, and want to have a bit-identical object file on re-compile, the solution is to add -frandom-seed="abcd" to your compile line (replace "abcd" with anything you want; it's common to use the filename as the value of random seed). Documentation here.
The reasons can be many:
Using macros like __DATE__ and __TIME__
Embedding counters that are incremented for each build (the Linux kernel does this)
Timestamps (or similarly variable quantities) embedded in the .comments ELF section. One example of a compiler that does this is the xlC compiler on AIX.
Different names as a result of name mangling (e.g. C++)
Changes in environment variables which are affecting the build process.
Compiler bug(s) (however unlikely)
To produce bit identical builds, you can use GCC's -frandom-seed parameter. There were situations where it could break things before GCC 4.3, but GCC now turns functions defined in anonymous namespaces into static symbols. However, you will always be safe if you compile each file using a different value for -frandom-seed, the simplest way being to use
the filename itself as the seed.
Finally I've found the answer!
c++filt command gave the original name of the function:
{unnamed namespace}: MinFunction(int, int)
In the source was:
namespace
{
MinFunction(int a, int b) { ... }
}
I named the namespace and got stable checksum of object file!
As I guess, the compiler didn't know how to name this namespace and used path to the source file plus some random number.
I need tool which is equivalent to:
$ echo 'char bar[] = {65, 66, 67};' >foo.c
$ gcc -c foo.c
I have a multi-megabyte binary file to be put to the bar array, and I need it without creating an .c file: I'd like the .o file be created directly from the binary file. Another option can be creating .s or .S files, but I'd like to avoid that as well. Is there a tool in binutils etc. which can do the job?
An update: gcc segfaults for a 9 MB binary file. as works, but it's slow and the the temporary .s file is too large.
You can use objcopy --add-section to create a section with contents found in a file. I think you'll need to use a linker script to add a symbol pointing at the start of the new section.
For a long time the easiest solution was creating an .s or an .S file.
binutils doesn't contain anything useful.
There is a 2-step trick which is fast and uses little memory:
Step 1. Create a and compile a .c file which contains the symbol of the right size, but it has a short signature instead of the real, long data. It should look like:
const char hi[1234567] = "SIGNATURE";
const char *hi_end = hi + sizeof(hi) / sizeof(char);
Step 2. Find the bytes SIGNATURE in the .o file generated by the compiler, and replace it (and the following '\0's) with the data from the real binary data file.
The Perl script cobjgen automates both steps. See this blog post for a more detailed analysis and usage instructions.
Is there a kind of "include" directive in RPM spec? I couldn't find an answer by googling.
Motivation: I have a RPM spec template which the build process modifies with the version, revision and other build-specific data. This is done by sed currently. I think it would be cleaner if the spec would #include a build-specific definitions file, which would be generated by the build process, so I don't need to search and replace in the spec.
If there is no include, is there an idiomatic way to do this (quite common, I believe) task?
Sufficiently recent versions of rpmbuild certainly do support %include:
%include common.inc
Unfortunately, they aren't very smart about it -- there is no known set of directories, in which it will look for the requested files, for example. But it is there and variables are expanded, for example:
%include %{_topdir}/Common/common.inc
RPM does not support includes.
I have solved similar problems with either m4 macro processor or by just concatenating parts of spec (when the "include" was at the beginning).
If you only need to pass a few variables at build time, and not include several lines from another file, you can run
rpmbuild --define 'myvar SOMEVALUE' -bb myspec.spec
and you can use %myvar in the spec.
I faced this same issue recently. I wanted to define multiple sub-packages that were similar, but each varied just slightly (they were language-specific RPMs). I didn't want to repeat the same boiler-plate stuff for each sub-package.
Here's a generic version of what I did:
%define foo_spec() %{expand:%(cat '%{myloc}/main-foo.spec')}
%{foo_spec bar}
%{foo_spec baz}
%{foo_spec qux}
The use of %{expand} ensures that %(cat) is only executed a single time, when the macro is defined. The content of the main-foo.spec file is then three times, and each time %1 in the main-foo.spec file expands to each of bar, baz and qux, in turn, allowing me to treat it as a template. You could easily expand this to more than one parameter, if you have the need (I did not).
For the underlying issue, there maybe two additional solutions that are present in all rpm versions that I am aware of.
Subpackages
macro and rpmrc files.
Subpackages
Another alternative (and perhaps the "RPM way") is to use sub-packages. Maximum RPM also has information and examples of subpackages.
I think the question is trying to structure something like,
two spec files; say rpm_debug.spec and rpm_production.spec
both use %include common.spec
debug and production could also be client and server, etc. For the examples of redefining a variable, each subpackage can have it's own list of variables.
Limitations
The main advantage of subpackages is that only one build takes place; This may also be a disadvantage. The debug and production example may highlight this. That can be worked around using strip to create variants or compiling twice with different output; perhaps using VPATH with Gnu Make). Having to compile large packages and then only have simple variations, like with/without developer info, like headers, static libraries, etc. can make you appreciate this approach.
Macros and Rpmrc
Subpackages don't solve the problem of structural defines that you wish for an entire rootfs hierarchy, or larger collection of RPMs. We have rpmbuild --showrc for this. You can have a large amount of variables and macros defined by altering rpmrc and macros when you run rpm and rpmbuild. From the man page,
rpmrc Configuration
/usr/lib/rpm/rpmrc
/usr/lib/rpm/redhat/rpmrc
/etc/rpmrc
~/.rpmrc
Macro Configuration
/usr/lib/rpm/macros
/usr/lib/rpm/redhat/macros
/etc/rpm/macros
~/.rpmmacros
I think these two features can solve all the problems that %include can. However, %include is a familiar concept and was probably added to make rpm more full-featured and developer friendly.
Which version are you talking about? I currently have %include filename.txt in my spec file and it seems to work just like the C #include directive.
> rpmbuild --version
RPM version 4.8.1
You can include the *.inc files from the SOURCES directory (%_sourcedir):
Source1: common.inc
%include %{SOURCE1}
In this way they will go automatically into SRPMS.
I've used scripts (name your favorite) to take a template and create the spec file from that. Also, the %files tag can import a file that is created by another process, e.g. Python's bdist-rpm.
This question already has answers here:
Is there a Linux equivalent of Windows' "resource files"?
(2 answers)
Closed 4 years ago.
I'm looking for a way to embed text files in my binaries (like windows resource system). I need something thats also platform independent (works in windows and linux). I found Qt resource management to be what I need but I'm not keen on my app depending on Qt for this alone. I also found this tool at http://www.taniwha.com/~paul/res/ .. but it is too platform specific.
The xxd utility can be used to create a C source file, containing your binary blobs as an array (with the -i command line option). You can compile that to an object which is linked into your executable.
xxd should be portable to most platforms.
If you're using QT 4.5, you can make sure that program is only dependent on one small piece of QT, such as libqtcore. QResource is a part of libqtcore.
You can simlpy append all kinds of data to your normal binary. Works in both Windows and Linux. You'll have to open your own binary at runtime and read the data from there.
However, I have to agree that embedding data in binaries is a strange idea. It's common practice to include such data as separate files packaged with the application.
That is not such a great idea. On Linux, for example, data is expected to be installed in a subdirectory of "$datadir" which is, by default, defined to be "$prefix/share", where "$prefix" is the installation prefix. On Mac OS X, resources are expected to be installed in $appbundle/Contents/Resources, where $appbundle is the name of the folder ending in ".app". On Windows, installing data in a folder that is a sibling of the executable is not an uncommon practice. You may be better off using the CMake build system, and using its CPack packaging features for installing/bundling in the default, preferred platform-specific manner.
Although bundling your resources into the executable, itself, may seem cool, it is actually a dangerous idea... for example, will the embedded data be allocated in an executable page? What will happen if you attempt to overwrite or modify the data? What if you want to tweak or modify the data at runtime? Things to think about.
This looks very promising: https://github.com/cyrilcode/embed-resource
CMake based and platform-independent.
As I also do not like the idea of converting files into C arrays only to have them converted back to binaries, I created my own resource compiler using LLVM and Clang:
https://github.com/nohajc/resman
I tested it on Windows, Linux and macOS but it can potentially be run on any platform supported by LLVM.
It is used like this:
Create header file, e.g. res_list.h
#pragma once
#include "resman.h"
// Define a global variable for each file
// It will be used to refer to the resource
constexpr resman::Resource<1> gRes1("resource_file1.jpg"); // resource with ID 1
constexpr resman::Resource<2> gRes2("resource_file2.txt"); // resource with ID 2
constexpr resman::Resource<3> gRes3("resource_file3.mp3"); // resource with ID 3
...
Run resource compiler
$ rescomp res_list.h -o res_bundle.o
Link res_bundle.o to your project
Use the resource files
#include "res_list.h"
...
resman::ResourceHandle handle{gRes1};
// ResourceHandle provides convenient interface to do things like:
// iterate over bytes
for (char c : handle) { ... }
// convert bytes to string
std::string str{handle.begin(), handle.end()};
// query size and id
unsigned size = handle.size();
unsigned id = handle.id();
The resource compiler parses res_list.h (using Clang) but instead of generating cpp files, it goes straight to the native object file (or static library) format (using LLVM).